1. Introduction
In recent years philosophers have spent considerable time scrutinizing the role of adaptive preferences in moral and political theory. While there appears to be a general consensus that adaptive preferences are problematic,Footnote 1 there is little agreement when it comes to explaining what it is exactly that makes them so problematic. Adaptive preferences are preferences shaped by the availability, or the perception of the availability, of options.Footnote 2 It is from this starting point that moral and political philosophers construct their explanations as to why adaptive preferences are problematic. These philosophers argue that adaptive preferences are problematic because they fail to correspond with objective values, undermine our autonomy, are the product of objectively bad circumstances, or some combination of these ideas.Footnote 3
These sorts of accounts have led Dale Dorsey to present a skeptical challenge against adaptive preferences.Footnote 4 He argues that adaptive preferences are a red herring in moral and political theory: what makes adaptive preferences problematic is not the fact that they are adaptive, but either that they fail to correspond with objective values, or that they undermine our autonomy, or that they arise from objectively bad circumstances, or so on. In defense of this skeptical conclusion Dorsey argues that it is possible to have adaptive preferences that do align with objective values, that are autonomous, or that are caused by objectively good circumstances.Footnote 5 If he is correct in arguing these points, as I think he is, then his arguments show that adaptive preferences are not inherently or necessarily problematic. This is as it should be. Many of our preferences are in fact adaptive, and they are not any worse for being so. The difficulty in dealing with adaptive preferences is that our intuitions concerning them seem to move in both directions – that some of our adaptive preferences are problematic while some are not. And insofar as adaptation is not inherently problematic in itself, any conceptual analysis of adaptive preferences and the accompanying explanation of their problematic nature ought to move beyond the bare fact of adaptation to available options by providing an account that allows us to distinguish between problematic instances of adaptation and non-problematic instances.
I intend to provide such an account. I argue that problematic instances of adaptive preference formation exhibit a peculiar error in reasoning and are consequently prima facie irrational. On the one hand, the very recognition of a distinction between problematic and non-problematic instances of adaptive preference formation allows my account to avoid the unqualified and therefore untenable claims that adaptive preferences fail to correspond with objective values, that they undermine our autonomy, or that they are the product of objectively bad circumstances. These sorts of accounts are too strong, which makes them easily susceptible to counterexamples, as Dorsey makes evident in his paper. Any conceptual analysis of adaptive preferences, if it is to be plausible, must make room for the idea that many of our adaptive preferences are not problematic.
My account will also move in some direction towards answering Dorsey's skeptical challenge by providing an explanation of the problematic nature of some adaptive preferences without entirely marginalizing the role of adaptation in that explanation. What makes some adaptive preferences problematic is that they exhibit a peculiar error in reasoning where individuals change the semantic content of the reasons supporting the new preference based only on the (un)availability of an option, and this error is a product – although not solely, as we will see – of adaptation itself. According to a distinction made by David Zimmerman, my view can best be understood as a hybrid account. Zimmerman draws a distinction between what he calls “psychological-process” accounts, which explain the problematic nature of some adaptive preferences by appealing to the psychological processes responsible for generating them, and what he calls “semantic-content” accounts, which appeal to the preferences’ semantic content.Footnote 6 There is an important connection between the two on my account: certain psychological phenomena cause a change in the semantic content of the reasons underpinning problematic adaptive preferences. Although my view incorporates psychological as well as semantic considerations in explaining the problematic nature of some adaptive preferences, it incorporates each in ways unlike what is currently on offer in the philosophical literature on adaptive preferences. In doing so, I believe it provides both a compelling and surprisingly flexible account of adaptive preference formation.
2. A promising start
The fact of the matter is that many of our preferences are adaptive, and this is not only not problematic but turns out to be quite a good thing. Consider the demands that morality or justice make on us as persons and citizens who possess the capacity to act morally. Morality and justice are often understood as representing the impartial, impersonal point of view, which is capable of conflicting with the partial, personal point of view. When these points of view conflict, as they often will, it is good that individuals adjust and adapt the pursuit of their self-interest according to the demands pressed by morality and justice. And besides the demands made by morality and justice, there are many other circumstances in our lives where we have to adjust and adapt our preferences either because of the actions of others or because of the limitations imposed by facts about the world and ourselves – the laws of physics, the current state of technology, the availability of resources, our genetic makeup, and so forth. It would indeed be very nice to fly around like Superman, but I would obviously need to adapt my preference to more realistic modes of travel.
What should be clear from these brief remarks is that any conceptual analysis of adaptive preferences that attempts to approach the concept with the assumption that adaptation is inherently problematic will quickly run afoul of the sundry instances in which adaptation is either unproblematic or quite beneficial.Footnote 7 Adaptation is a persistent feature of individual psychology and preference formation in a wide number of circumstances.
In response to these points, one might have the inclination simply to make the requisite distinction between problematic and non-problematic instances of adaptation, and then proceed to explain the former by appealing to one of the explanations previously enumerated. While these sorts of responses certainly accommodate the idea that adaptation is not inherently problematic, they do so only by succumbing to the skeptical challenge pressed by Dorsey – adaptation becomes explanatorily irrelevant in explaining why those preferences are problematic.
What we need is an account of adaptive preference formation that accommodates the thought that only some of our adaptive preferences are even prima facie problematic, and the explanation for why those adaptive preferences are problematic should not reduce to an appeal to some further normative consideration. In this section I want to consider one account of adaptive preference formation that I believe points us in the right direction but ultimately falls short. This account is offered by Luc Bovens.Footnote 8
Bovens, in following Jon Elster, makes it clear that he recognizes that adaptation is a persistent feature of individual preference formation. The question then is what leads us to evaluate cases of adaptive preference formation differently. Why do we find one problematic, but not the other? Despite the fact that I believe Bovens ultimately misidentifies the crucial feature that makes some adaptive preferences problematic, I do think that he provides an illuminating framework for conceptualizing adaptive preferences – one that I plan to utilize.
In his analysis of adaptive preferences, Bovens adopts the account of practical reasoning set out by Donald Davidson in his famous paper on weakness of the will.Footnote 9 According to the Davidsonian account of practical reasoning, for most, if not all, potential instances of human action there will be conflicting evaluative reasons that support either performing that potential action or performing some other alternative. I may be trying to decide, for example, whether I ought to take a month-long vacation in Italy or Spain. There will be a number of separate evaluative reasons – what Bovens calls “criterial judgments” – that favor a vacation to Italy over a vacation to Spain. Along one criterion, say, the opportunity to enjoy masterful paintings, my criterial judgment might identify the vacation to Italy as being superior. Other criteria, however, may provide reasons supportive of taking a vacation to Spain instead of Italy. Now according to Davidson, both my decision and the resulting action should be the product of my all-things-considered judgment (AJ). My AJ adjudicates over all the relevant and conflicting criterial judgments, as well as other considerations, by ranking the importance of each in order to reach a determinate conclusion as to what I should prefer and pursue. So while it may be true that the vacation to Italy wins along some criteria, it turns out that the vacation to Spain wins along more criteria or more important criteria. As a result, my AJ forms a conclusive judgment in favor of a vacation to Spain. Importantly, for both Davidson and Bovens an individual's AJ establishes a test for determining the rationality of a preference: a preference is rational only when it conforms to the AJ.Footnote 10
Bovens proceeds to apply this framework to adaptive preference formation in order to distinguish non-problematic instances of adaptive preferences from problematic ones. What makes an adaptive preference rational rather than irrational, according to Bovens, is its conformity with an individual's AJ.Footnote 11 A rational adaptive preference is the product of alterations to criterial judgments – either on the basis of a new evaluation of relevant properties or in the relative ranking of each criterial judgment – that alter in turn an individual's corresponding AJ. This means that a newly formed adaptive preference must be supported with changes to the criterial judgments that form the basis of the AJ, so that it remains in conformity with the AJ. Problematic adaptive preferences, in contrast, are irrational precisely because an individual does not make any alterations to the criterial judgments that form her AJ. Because her AJ remains the same, her new adaptive preference fails to conform to it.
Consider how Bovens’ analysis explains the irrational adaptive preference formation of the fox from La Fontaine's fable. The fox prefers to eat the grapes hanging from a branch, but upon finding the grapes to be unreachable, she tells herself that she did not really want them anyway. According to Bovens, this adaptive change in the fox's preference is irrational because it is not accompanied by any changes in her criterial judgments that would justify a change in her AJ. With no change to her AJ, her adaptive preference does not conform to it, and that makes her adaptive preference irrational.
3. The irrationality of some adaptive preferences
By applying Davidson's account of practical reasoning to adaptive preference formation Bovens provides a constructive and accessible framework for conceptualizing adaptive preferences. There are, however, two essential revisions that must be made to Bovens’ analysis. First, I believe that Bovens gets the distinction between problematic and non-problematic adaptive preferences backwards to a certain extent: contra Bovens I want to suggest that there are no changes (at least initially) in the criterial judgments regarding the unavailable (or available) option when individuals form non-problematic adaptive preferences and that there are changes in the criterial judgments regarding the unavailable (or available) option when individuals form problematic adaptive preferences. The general idea here is that an individual's determination of an option's unavailability causes her to negatively re-evaluate that option. This re-evaluation is not simply a matter of repositioning the relative ranking of different criterial judgments by adding additional weight to one criterial judgment rather than another, though the problem may manifest itself in this manner as well.Footnote 12 To be more specific, the unavailability of an option causes some individuals to exhibit a peculiar error in reasoning where they change the very semantic content of the criterial judgments that form the basis of their new adaptive preference. The problem, of course, is that there is no reason to think that the (un)availability of an option should alter the qualitative properties of the objects, actions, or outcomes which constitute that option. These are obviously logically distinct considerations; as a consequence, the same semantic changes will not be reflected in the criterial judgments that form these individuals’ AJs, where that error in reasoning is absent. The end result is a discrepancy between individuals’ adaptive preferences and their AJs, which makes them prima facie irrational.
The second necessary revision that must be made to Bovens’ analysis concerns the status of problematic adaptive preferences. As we saw in the previous section, he classifies problematic adaptive preferences as irrational because they do not conform to an individual's AJ. An individual's newly formed adaptive preference is not based on any changes to the criterial judgments underpinning her AJ, and the result is a discrepancy between the latter and her adaptive preference. This discrepancy between the two renders the adaptive preference irrational.
In contrast to Bovens’ analysis and according to my account, problematic adaptive preferences are prima facie irrational, and while this is a relatively weaker attribution of irrationality, this weaker attribution better accords with plausible views of practical reasoning and affords my account greater explanatory flexibility. To say that problematic adaptive preferences are prima facie irrational is to say that there is an initial presumption in favor of treating those preferences as irrational – an initial presumption which may be overturned on further inspection. How could this be possible? Recall that on the Davidsonian view of practical reasoning, rationality is a matter of conformity between an individual's preference and her AJ. The problem with Bovens’ analysis is that it appears to treat criterial judgments as if they exhaust the considerations relevant to the formation of an AJ. Any plausible account of practical reasoning, however, cannot reduce the bases of an AJ only to criterial judgments. Recall that criterial judgments, according to Bovens, represent different evaluative criteria according to which an individual will form a preference. These criteria are naturally specified in a manner relative to the nature of the objects, actions, or outcomes being compared: if an individual is trying to decide which of two automobiles to purchase then the type of criterial judgments will be determined by criteria relevant to automobiles, such as color, year, mileage, make, and so on. But under any plausible view of practical reasoning, including Davidson's view, these criterial judgments do not exhaust the considerations relevant to forming an all-things-considered judgment. There are reasons independent of evaluative criteria that will also play an indispensable role in determining an AJ. There may be more general questions and considerations regarding morality, of long-term plans, the perceptions of others, feasibility concerns, questions about whether and to what extent pursuing an option will contribute to one's happiness, whether an option aligns with one's values and self-identity, and so on, that all play an independent yet essential role in determining an AJ.Footnote 13 Such a judgment reflects the idea that our preferences and actions are not based solely on an estimation and ranking of the qualitative properties of different objects or actions that comprise our options. Where it is problematic to allow concerns about availability to change our original estimation of the qualitative properties of our options, it is not problematic to allow concerns about feasibility to guide our preferences and actions. After all, the prospect of success in pursuing a particular option is an appropriate consideration that should factor in the formulation of an AJ, for it would matter little how ideal one option may be according to all relevant criteria if there is no possibility of ever achieving it.Footnote 14
Once we recognize that an AJ need not depend exclusively on criterial judgments, we can better understand how an individual may form a problematic adaptive preference that is prima facie irrational, which turns out to be, on further inspection, rational. The unavailability of a particular option may lead an individual to change a criterial judgment regarding that option in a problematic manner, but insofar as her adaptive preference turns out to promote her self-interest, as just one example, it may nonetheless conform to her AJ. As long as considerations other than criterial judgments can play a role in determining an individual's AJ, then it is possible for individuals to form rational adaptive preferences, even when part of their reasoning for doing so is problematic.
My reason for emphasizing this point of difference between our two accounts is that this difference has important implications for how we understand adaptive preference formation under oppressive conditions. In many oppressive circumstances, it is often in conformity with individuals’ AJs to form self-demeaning adaptive preferences that also happen to promote their self-interest. But as long as their adaptive preferences conform to their AJs in these cases, then it is not clear on what basis Bovens could label them as problematic. However, insofar as my account explicitly distinguishes between the various sorts of considerations that may all play a role in determining an individual's AJ, then it allows us to see how an individual may form a problematic adaptive preference that turns out to be in her self-interest, all things considered. By separating the explanation of what makes some adaptive preferences problematic from the consequences of holding those preferences in various real-world scenarios, my account provides better insight and guidance as to how both problematic and non-problematic adaptive preferences should be addressed in non-ideal circumstances. I will return to these points in more detail in the fourth section where I apply my analysis to several concrete examples.
Thus far I have identified and outlined, in an admittedly abstract manner, two revisions that must be made to Bovens’ analysis. I want to return to the first proposed revision, as it captures the fundamental idea behind my analysis of adaptive preference formation. I introduced in abstract terms the claim that what makes some adaptive preferences problematic is when individuals negatively re-evaluate an unavailable option by changing the semantic content of the relevant criterial judgments. To better understand this change and why it is problematic, let us return to the fable of the fox and the grapes.
As the story goes, the fox prefers to eat the grapes hanging from the branch, but upon finding the grapes to be unreachable, she convinces herself that she never really wanted the grapes in the first place. How does she explain to herself what appears to be a sudden change in her preferences? Her explanation, of course, is that the grapes are sour. Now in order to make sense of what is happening here it is necessary to recognize the possibility of describing her behavior (and her beliefs) in two divergent ways, depending on which perspective we adopt. From the first-person perspective, from the fox's own perspective, there may appear to be no preference change at all. The fox believes, or at least has convinced herself that she believes, that the grapes have always been sour. It is this persistent belief in the grapes’ sourness that justifies her equally persistent preference, from her perspective, for not eating the grapes.
It is at this point that the description from the first-person perspective diverges from the description offered from the third-person perspective. From a third-person perspective, from an observer's perspective, it is obvious that the fox's preference for not eating the grapes is an adaptive response to an unavailable option, and that her belief in the grapes’ sourness is a rationalization designed to reduce her frustration. That this is indeed the correct description is clear insofar as the fox's initial reach for the grapes belies the very rationalization she offers only once she discovers her inability to secure them. After all, if sourness were a sufficient reason for preferring to not eat the grapes as she now claims it is, and her belief in the grapes’ sourness never changed, there would have been no reason to reach for the grapes in the first place. That her previous action belies her proffered explanation for her (new) preference shows that a change has occurred within her criterial judgments, although one she would not be likely to acknowledge, for doing so would threaten to undermine the very plausibility of her rationalization. Despite what she now claims, her evaluation of the grapes according to the criterion of flavor has indeed changed: whereas she believed the grapes to be juicy before, she now believes them to be sour, and her new adaptive preference is based on the change to this criterial judgment. Of course, what makes this change problematic is that it is not based on any direct experience (or other relevant new information) of the flavor of the grapes, for she has not actually tasted them in order to revise her criterial judgment. Rather, the change is based entirely on her inability to acquire the grapes. The unavailability of the grapes has led her to change a criterial judgment regarding them in order to mollify her frustration.
Whereas problematic adaptive preferences exhibit this error in reasoning, non-problematic adaptive preferences are distinguished by its absence. Consider the fox once more. In this version she reaches for the grapes, but upon discovering her inability to reach them she offers no rationalization that alters any of her criterial judgments regarding the taste of the grapes; she does not tell herself that the grapes are sour. On the contrary, she continues to acknowledge the desirability of possessing and eating the grapes, but she nonetheless alters her preference to prefer not eating the grapes or to prefer some other edible and available option. Her new preference is clearly adaptive, but it does not exhibit the same problematic reasoning evident in the first example.
At this point Bovens would likely respond that without some change in the fox's criterial judgments that would justify altering her AJ her adaptive preference – in this second case where her reasoning does not exhibit the same problem as in the first case – would nonetheless fail to conform to her AJ, thereby rendering her adaptive preference irrational. There are two points that must be made in response to this potential rebuttal by Bovens. These points should help to further clarify the basis for the distinction between problematic and non-problematic adaptive preferences on my account.
It was mentioned previously that it is best to understand an individual's AJ as adjudicating not only over criterial judgments, as Bovens appears to suggest, but also independent considerations, such as the feasibility of an option and the likelihood of success in pursuing an option. As these sorts of considerations can and often will play an important role in the determination of an individual's AJ, it is possible for an individual to alter her AJ without changing any criterial judgments. Not only is it an error of reasoning to conclude that the unavailability of an option somehow alters its qualitative properties – as happens in the formation of problematic adaptive preferences – but it accords with common sense to think it possible for individuals to continue to acknowledge the merits of a particular option, while acknowledging at the very same time that it would not be best, all things considered, to pursue that option. This is precisely what occurs in the formation of non-problematic adaptive preferences: the infeasibility of an option leads an individual to alter her AJ, which then serves as the basis for her newly formed adaptive preference. In doing so, she continues to acknowledge the merits of the unavailable option, but she also concedes that because it is not possible to achieve that option it would not be best, all things considered, to pursue it any further. And in this manner she forms an adaptive preference that conforms to her AJ without altering any criterial judgments.
Furthermore, we have seen that Bovens identifies the problematic nature of adaptive preferences in terms of a discrepancy between an individual's adaptive preference and her AJ. The division between problematic and non-problematic adaptive preferences is thus demarcated by the ideas of rationality and conformity: adaptive preferences are not problematic insofar as they are rational – that is, insofar as they conform to an individual's AJ. Although I explained in the previous paragraph how it is possible for my account to accommodate the idea that some adaptive preferences may be rational without any alterations to criterial judgments, it is important to underscore that the distinction between problematic and non-problematic adaptive preferences is not, on my account, a matter of whether the preferences in question are irrational or rational. I have indeed claimed that problematic adaptive preferences are prima facie irrational, but it does not follow, on my account, that non-problematic adaptive preferences are necessarily prima facie or fully rational – that is, that they must conform to an individual's AJ.
This point may seem obscure at first. But because an individual's AJ comprises a number of different considerations, determining whether an adaptive preference is rational will be more complicated than one might initially suspect. Just as it is possible to have problematic adaptive preferences that turn out to be rational on further inspection, it is also possible to have non-problematic adaptive preferences that are irrational. Adaptive preferences are preferences shaped by the availability, or perceived availability, of options. There is no reason to assume, however, that an individual's pre-adaptive preference must always be rational, that it must always conform to her AJ. For example, Jane may have a preference to rob a bank over robbing a gas station (with all other options far behind in her rankings). The demands of morality and long-term considerations of self-interest converge to form an AJ staunchly in opposition to either of those two preferences, however. Upon further consideration, Jane comes to believe that a bank heist is not a feasible option, and she forms an adaptive preference in favor of robbing a gas station instead. While the relative feasibility of robbing a gas station, compared to robbing a bank, appropriately plays a role in the determination of Jane's AJ, it would not outweigh the countervailing considerations of morality and long-term self-interest, and the result is that her AJ continues to adjudicate against robbing a gas station as the best option. Whether Jane has formed a problematic or non-problematic adaptive preference depends on whether she has let the unavailability of robbing a bank change her criterial judgments regarding that option. Perhaps she continues to acknowledge the superiority of robbing a bank compared to robbing a gas station, yet her reasoning appropriately allows the infeasibility of doing so to guide her change in preferences. Insofar as there is nothing inherently problematic in adapting our preferences to our circumstances, Jane does not commit the error in reasoning I have specified in forming an adaptive preference to rob a gas station instead of a bank. Her adaptive preference is not therefore problematic on my account, but it is irrational insofar as it fails to conform to her AJ. And if we think her preference problematic, then it is problematic because of its irrationality (and immorality), not because it is adaptive.
The upshot of this example is that we cannot distinguish between problematic and non-problematic adaptive preferences based on their irrationality or rationality. Some may think that this is a weakness of my account, but I would insist that this explanatory flexibility underscores its strength. Depending on the circumstances and the content of the preferences in question, non-problematic adaptive preferences may be either rational or irrational. Adaptive preferences that are problematic, in contrast, are distinguished by the error in reasoning that I have specified – they are distinguished by the fact that individuals reason in such a manner where the unavailability of an option alters that option's qualitative properties from their perspective. Problematic adaptive preferences are thus prima facie irrational because there is no reason to think that the unavailability of an option should alter its properties in any way. As this error signals a problem in the individual's reasoning, then it is a problem that persists regardless of what an individual's circumstances are. However, once we take into account an individual's particular circumstances, it is possible even for a problematic adaptive preference to conform to an individual's AJ, despite that initial error in reasoning. This is not too uncommon an occurrence for individuals in oppressive circumstances, and I will turn to several real-world examples in the following section to illustrate its plausibility.
It is worth reiterating that the explanation behind the distinction between problematic and non-problematic adaptive preferences must be severed from any reference to the sundry effects and consequences that arise from holding adaptive preferences. Any explanation that conflates the two will fall prey to Dorsey's skeptical challenge: not only will some adaptive preferences generate harmful consequences, but some non-adaptive preferences will also generate those same consequences. By tying the problematic nature of some adaptive preferences to the various consequences that come about from holding those preferences – whether those consequences are specified in terms of rationality, autonomy, well-being, or some other metric – the phenomenon of adaptation becomes, as Dorsey points out, a red herring. What draws our attention is not whether the preferences in question are adaptive but whether they generate the relevant sort of consequences. Insofar as my account of the problematic nature of adaptive preferences locates the problem in the way that individuals view and reason about the unavailable options, it not only provides a compelling conceptualization of adaptive preferences, but it provides an account capable of responding to Dorsey's skeptical challenge.
Furthermore, the error in reasoning I have identified does not apply exclusively to criterial judgments about an unavailable option; the same error may infiltrate the manner in which individuals evaluate an available option. They may change the criterial judgments underpinning an initially less desirable but available option in a more positive manner in order to achieve similar psychological ends. Although this version presents an equally problematic instance of adaptive preference formation, it does need to be distinguished from cases in which individuals come to genuinely appreciate what a newly formed adaptive preference entails.Footnote 15
To see why this distinction is important consider the following two examples. I prefer to be a professional basketball player over a professional engineer. Unfortunately, given that my relative height renders my prospects of ever becoming a professional basketball player fanciful, I opt to become a professional engineer instead. In making this choice I adjust my AJ in light of the infeasibility of becoming a professional basketball player and form an adaptive preference in conformity with that judgment. Though I have altered my AJ, I have not altered any of my criterial judgments regarding either profession. Yet, after some time working as an engineer, I come to better appreciate the profession and what it involves. As a result, I change the content of the corresponding criterial judgments to the extent that my AJ now offers a more holistic and conclusive ruling in favor of engineering. Where some of those criterial judgments may have previously given the edge to becoming a professional basketball player, they now speak in favor of a career in engineering. Perhaps, for instance, I have come to appreciate not having to constantly travel as is typically required of professional basketball players, whereas before this seemed to be a distinct benefit in favor of the latter. Initially, all else being equal, I would have preferred a professional career in basketball over a professional career in engineering. The infeasibility of the former, however, led me to adjust my preference in favor of the latter, and in doing so, I eventually came to appreciate what engineering had to offer. The subsequent changes to my criterial judgments were the product, not of the unavailability of a professional career in basketball, but of my directly relevant experiences as a professional engineer (and perhaps of other relevant changes in my life).
Now compare this example to another version of La Fontaine's fox. In this version of the fable, the fox is also quite hungry. Unfortunately, all that is available is a batch of putrid grapes. If she could, she would certainly prefer to eat anything else over the rotting grapes. Because of their availability and her gnawing hunger, however, she convinces herself that the grapes are quite juicy and forms an adaptive preference in favor of eating them. She changes her criterial judgments regarding the qualitative properties of an option simply because of its availability.Footnote 16 And this is what makes it problematic. In the traditional version of the fable, the fox prefers to eat the grapes, but cannot, so she negatively changes a criterial judgment to support her adaptive preference; in this version of the fable, she prefers to not eat the grapes, but must, so she positively changes a criterial judgment to support her adaptive preference.
It should be clear by now that what distinguishes problematic from non-problematic adaptive preferences is that the former involve a change to a criterial judgment based only on the (un)availability of an option. The (un)availability of an option causes a change to the semantic content of a criterial judgment. But because there is no reason to think that the (un)availability of an option provides any basis for an alteration of its qualitative properties (or in the weights attached to criterial judgments), no parallel change ought to occur in the AJ. As a consequence, the fox's newly formed adaptive preference fails to conform to her AJ and is therefore prima facie irrational.
The distinction between problematic and non-problematic adaptive preferences is easier to observe if individuals who form adaptive preferences are presented with an explicitly hypothetical question. If the previously preferred but unavailable option were suddenly to become available, would the individual now prefer it? Individuals with problematic and non-problematic adaptive preferences will answer this question differently. Individuals with non-problematic adaptive preferences will typically respond that they would prefer the original option, if it were to become available. (I say typically because we have to allow room for circumstances in which individuals come to prefer the second option, rather than the original option, by adjusting their criterial judgments in light of their experiences, as in the engineer example.) The reason that individuals with non-problematic adaptive preferences will respond in this way is because the unavailability of the original option did not cause them to change (negatively or positively) a criterial judgment underpinning either the unavailable or the available option. They continue to recognize the worth or value of the unavailable option, such that if it were to become available, they would select it. In contrast, individuals with problematic adaptive preferences will answer the hypothetical question negatively precisely because the unavailability of the original option did cause them to change (negatively or positively) a criterial judgment underpinning either the unavailable or the available option. And since the question is only hypothetical, an affirmative answer would only threaten to undermine the very success of their rationalization, which depends on that change to the criterial judgment supporting the new adaptive preference.
4. Applying the analysis
It is a concern for adaptive preferences formed in real-world circumstances – such as when individuals are forced to confront and navigate the effects of oppression in its psychological, social, and economic forms – and not for foxes from fanciful fables, that has been primarily responsible for the heightened attention moral and political philosophers have directed towards adaptive preferences. Any analysis of adaptive preferences that holds little plausibility or provides little guidance for understanding and addressing adaptive preference formation in real-world circumstances might reasonably be seen as possessing marginal value. For this reason, I intend to apply my analysis to some concrete examples found in the literature to better assess its overall plausibility. The two examples I discuss – which are taken from Serene Khader's book, Adaptive Preferences and Women's Empowerment Footnote 17 – are meant to illustrate some of the epistemological difficulties involved in diagnosing and treating individuals’ adaptive preferences in real-world circumstances. I take it that any plausible analysis of adaptive preference formation should provide the right sort of explanation in both of these examples – the sort of explanation that does not apply problematic adaptive preferences indiscriminately to individuals without also managing an appropriate sensitivity to the features that constitute and complicate their objective circumstances.
4.1. Psychologizing the structural
One sort of epistemological difficulty Khader identifies – what she calls “psychologizing the structural” – is the mistaken tendency to treat individuals’ beliefs, attitudes, and values, rather than their circumstances, as the proximate cause for their adaptive preferences.Footnote 18 As one example, Khader considers cases of women who choose not to pursue their own literacy. At first glance, the preferences held by these women might suggest a lack of appreciation for education. As a consequence, some development practitioners – those engaged in international development who address oppression in the field – as well as philosophers may misdiagnose the source of the problem based on these women's expressed preferences: they might conclude erroneously that cultural and social circumstances have led these women to not appreciate the value of education, and they might consequently recommend inappropriate solutions directed at altering the beliefs and values held by these women.
What makes these “lower-order” adaptive preferences deceptive from Khader's perspective is that they may reflect rational responses to difficult circumstances in which resources for education are limited, rather than a lack of appreciation for the value of education. Because education requires a significant financial investment and these women are constrained by their meager economic resources, they choose to invest their limited resources in the education of their children, who will provide greater long-term economic return. Their adaptive preferences do not reflect misguided beliefs about the value of education, but they are rational responses to straitened economic circumstances.
At the end of the third section I proposed one method for identifying non-problematic and problematic adaptive preferences. What are individuals’ responses to the question in which the unavailable option is hypothetically made available? Those who have formed non-problematic adaptive preferences will provide a different response to this question compared to those who have formed problematic adaptive preferences. Some of the women in Khader's example will profess their preference for their own education, and this will show that the problem in their case is not their adaptive preference to not pursue their own education but their economic circumstances, which preclude the possibility of educating their children and themselves at the same time. Because these women do not exhibit problematic adaptive preferences, the appropriate solution in this case is to address those socioeconomic conditions responsible for producing their adaptive preferences. Still, there may be other women from Khader's example, who in response to the question, continue to profess a preference that demonstrates a genuine lack of appreciation for the value of education. These women have come – consciously or unconsciously – to disvalue an education for themselves because that option is not available.Footnote 19 The proper recourse for action from a development practitioner's perspective in this case ought to be to challenge the new criterial judgments that support the adaptive preference, in addition to improving the prevalent socioeconomic conditions.
In the third section of the article I stressed that on my account problematic adaptive preferences are prima facie irrational. Because an individual's AJ does not adjudicate only over criterial judgments, it is possible for an individual to form a problematic adaptive preference that turns out to be rational. Women who form a problematic adaptive preference for not pursuing their own education demonstrate how this is possible. These women form a problematic adaptive preference by changing their criterial judgments regarding the value of an education because that option is unavailable, and in doing so, they form an adaptive preference that initially fails to conform to their AJ. Further consideration of the facts that constitute their objective circumstances – such as the fact that an education for their children will produce greater economic returns and will thereby improve the well-being of the entire family – will, however, rightly play a role in the formation of their AJ. The effect of considerations like these allows even a problematic adaptive preference to conform to an individual's AJ, thereby rendering that adaptive preference rational.
Given the circumstances these women find themselves in, it is quite rational for them to form such an adaptive preference, but we must avoid conflating, as Bovens does, the rationality of holding an adaptive preference in a given set of circumstances with the innocuous nature of that preference. Even though individuals may hold the very same adaptive preference, the reasoning and justification behind that adaptive preference may be very different from one individual to another. It is only by being sensitive to this point that development practitioners will be able to respond appropriately to adaptive preferences in real-world circumstances.
4.2. Misidentifying imposed trade-offs
A second epistemological difficulty involved in diagnosing adaptive preferences includes circumstances where individuals have to choose which components of well-being to satisfy when it is not possible to achieve them all simultaneously. Khader's discussion of this epistemological difficulty focuses on the case of a woman who engages in condomless sex with her husband to avoid angering him.Footnote 20 This woman's choice to engage in condomless sex with her husband may or may not reflect “inappropriately adaptive attitudes towards her health.” Khader defines inappropriate adaptive preferences as preferences that are both inconsistent with human well-being (which is objectively determined) and formed under conditions non-conducive to human well-being. What makes the status of certain preferences ambiguous on Khader's account is that an individual may prefer one option, which is itself incompatible with her well-being, in order to preserve her access to goods that are conducive to her well-being. A woman may thus choose to engage in condomless sex, even though doing so poses a risk to her bodily health, to avoid angering her husband whom she depends on economically and socially. If a development practitioner were to take this woman's preference only at face value, she might propose an ill-suited remedy, such as providing information designed to alter the woman's beliefs regarding the health risks to which she is vulnerable by having condomless sex. On the other hand, it may be the case that this woman's preference – instead of reflecting a preference for incorrect values – reflects an attempt to navigate a difficult situation where a trade-off between two components of well-being is unavoidable. Because it is not possible for her to achieve both bodily health and economic well-being, she chooses to risk her bodily health for her economic well-being. And if this were the correct motivation for her preference, then providing her with additional information regarding the health risks of condomless sex would do little to alleviate and improve her situation or to alter her adaptive preference.
We have seen that on Khader's account what makes a preference inappropriately adaptive is its failure to align with well-being and when it is the product of conditions that are non-conducive to well-being. What accounts for the ambiguous status of this woman's preference is that it sacrifices one component of well-being for another. A superficial evaluation of her preference would lead one to conclude that it satisfies both conditions and is therefore inappropriately adaptive. A more detailed inspection of her preference, however, would show that her preference actually does align with her well-being, given the circumstances, so that the first condition is not satisfied, and her preference is not, as a result, inappropriately adaptive. The problem with this sort of explanation is that both of the conditions enumerated by Khader have little to do with adaptation. Preferences are problematic not because they are adaptive but because they fail to align with well-being and because they stem from conditions non-conducive to well-being. The fact that the preferences in question are adaptive is a red herring.Footnote 21
What sort of answer does my analysis provide in this case? Once we identify an individual as having formed an adaptive preference, my analysis proceeds to an evaluation of the individual's evaluative beliefs regarding both the unavailable option and the available option to determine whether the adaptive preference in question is problematic. Has a woman with an adaptive preference for risking her bodily health in sexual matters come to disvalue her bodily health altogether? Perhaps she continues to monitor her bodily health in other ways and in other areas of her life, or conversely, perhaps a belief regarding the unimportance of her bodily health in sexual matters has led to a complete disregard of her bodily health in other ways and in other areas of her life. Her beliefs and behavior in these other areas will allow us to determine whether her adaptive preference is indeed problematic. As long as she has not changed her criterial judgments regarding the desirability of bodily health simply because it may be no longer possible for her to maintain it, then she would not possess a problematic adaptive preference. A woman who has formed an adaptive preference for condomless sex for the sake of preserving her access to economic and social goods will not harbor the same sort of negative views towards her bodily health (or sex with a condom, or more positive views either towards bodily harm or condomless sex) compared to a woman who has formed a problematic adaptive preference. If the option to engage in sexual acts with a condom were suddenly to become available, then a woman with a non-problematic adaptive preference would not continue to engage in condomless sex, as her beliefs and judgments regarding the merits of that option have not changed. That her adaptive preference for condomless sex may be incompatible with the components of human well-being provides perhaps an additional reason for viewing it as problematic, but the primary reason for labeling it as problematic, if indeed it is, is the re-evaluation of an option's qualitative properties simply in virtue of its (un)availability.
5. Answering the skeptical challenge: are adaptive preferences a red herring?
I believe the account I have offered in this article goes some way in providing a constructive response to the skeptical challenge without fully answering it. In contrast to many of the typical accounts of adaptive preferences, my analysis is psychological and semantic. Adaptation appears to satisfy an important psychological need generated in response to the generally recalcitrant nature of the world. Human beings appear to have a need to realize their wants and preferences in the world. We do not foster wants and formulate preferences merely to engage in an exercise of imagination; we seek to achieve harmony between what goes on in our minds and what goes on in the external world. Due to many intransigent features of the world – the actions of others, the state of technology, the availability of resources, our own natural limitations, and so on and so forth – it is not always possible to actualize our preferences. One response to this failure may be an attempt to exert our will on the world in a domineering fashion: we may attempt to shape the world – through force or other means – to fit our preferences. But for many people and for many reasons this will not be possible. When it is not possible to exert our will on the world, a second response is to adapt our preferences to fit the world. By adapting our preferences, we at least reduce, and in some cases may entirely eradicate, the psychological tension and anxiety that are symptoms of a disharmony between the way we want the world to be and the way it actually is. Insofar as this sort of disharmony is an inevitable and persistent feature of human life and insofar as adaptation is capable of reducing many of the psychological ailments that may arise from it, then adaptation appears to be a healthy and perhaps essential feature of human psychology.Footnote 22
Though all this likely strikes the reader as obvious, there is, I believe, a prevailing prejudice to assume that the sort of psychological harmony mentioned above depends on a re-evaluation of either the unavailable option or the available option. How else would an individual achieve psychological harmony unless she convinces herself of the inferiority of an unavailable option? Would she not be living in a state of perpetual psychological tension without this re-evaluation? I am not so sure, or at least, I am not sure that it must always pose significant psychological problems if true. It is possible, both conceptually and empirically, for individuals to form adaptive preferences without changing their criterial judgments by taking into consideration other relevant factors. As these individuals continue to recognize the merits of the unavailable option, then they may never reduce that psychological tension altogether. But, for reasons I touch on below, they may not need to.
Unlike other accounts of adaptive preferences, which tend to explain their problematic nature by appealing to some other normative consideration in a way that renders adaptation irrelevant, my account locates the problem in adaptation itself. The error in reasoning that I have identified is caused both by adaptation and by a particular form of adaptation; this error cannot, in other words, be separated from the process and the act of adapting our preferences to our circumstances. This error evinces a very particular psychological response by some individuals to a necessity imposed by the external world. As not all individuals will share this exact psychological response to the ubiquitous necessity of having to adapt our preferences – some will, after all, form non-problematic adaptive preferences – one might wonder what causes the difference in these psychological responses. What causes one individual, but not another, to reason in this manner and to form a problematic adaptive preference?
It is at this point that I must concede the limited extent to which my account satisfactorily answers Dorsey's skeptical challenge. I have distinguished between problematic and non-problematic adaptive preferences based on the manifestation of a peculiar error in reasoning. Whereas problematic adaptive preferences exhibit this error, non-problematic adaptive preferences are distinguished by its absence. The plausibility of drawing a conceptual distinction between problematic and non-problematic adaptive preferences implies that the property of adaptation does not tell us the entire story. We know that this problem occurs when individuals have to adapt their preferences, but I can offer no explanation here as to why some individuals respond by adapting their preferences in one manner, such that they exhibit the error in reasoning I have identified, while others manage to adapt their preferences in another manner. In fact, the answer to this inquiry is not one I believe philosophers can provide. We must turn instead to psychologists to explain why there are such different responses to the same stimulus, like the need to adapt one's preferences to one's circumstances.
Although this may strike the reader as an unsatisfactory conclusion, some progress has been made in responding to Dorsey's skeptical challenge. We know that because not all individuals form problematic adaptive preferences, there must be some psychological phenomena, in addition to the need to adapt one's preferences, which play an essential role in explaining why some individuals form problematic adaptive preferences while others do not.
Although there may be a number of psychological phenomena in addition to adaptation, and we may even speculate that none of these other phenomena is exclusive to the process of adaptation, it certainly seems relevant from an explanatory standpoint that adaptation is one necessary cause within that overall explanation. Perhaps a lack of self-esteem, or optimism, or resiliency may, for instance, explain an individual's inability to pursue one of her quite feasible goals at a given moment in time. But it may be the case that only when a low level of self-esteem or a high level of pessimism is combined with the need to adapt one's preferences that an individual then exhibits the error in reasoning I have identified. Perhaps it is the presence of the more positive versions of these psychological traits that explains why some individuals are better able to cope with some level of psychological tension and are able to form, as a consequence, adaptive preferences that do not rely on the same sort of erroneous reasoning.
The answer to this inquiry, whatever it may be, would indeed provide an informative explanation because it would show that a particular effect – an error in reasoning where individuals change the content of criterial judgments (or their rankings) based only on the (un)availability of an option – can be caused only by a particular set of causes; and that answer would also show that adaptation is a necessary, but non-sufficient, cause within that set. Just as when two chemicals, which may have beneficial effects when taken on their own, have a more inimical effect when combined, the combination of adaptation with other psychological phenomena is responsible for a peculiar error in reasoning. As of now, we have an incomplete psychological explanation as to why this problem occurs, but while the explanation is currently incomplete, we also know that adaptation is not a red herring, for it plays an essential role within that explanation.Footnote 23
